Remodeling Numerical Representation for Text Generation on Small Corpus


Tan, Aristotle and Goh, Hui Ngo and Wong, Lai Kuan (2019) Remodeling Numerical Representation for Text Generation on Small Corpus. In: 2019 2nd International Conference on Machine Learning and Natural Language Processing, 18-22 Dec. 2019, Sanya, China.

[img] Text
154.pdf - Published Version
Restricted to Repository staff only

Download (852kB)


Data-to-text generation aims to generate natural language descriptions from non-linguistic data. Recent research on data-totext generation often uses a neural encoder-decoder architecture due to its simplicity to work across multiple domain. In this study, we aim to investigate two input encoding strategies: (1) numeral encoding as baseline, and (2) numeral as sequence-of-character tokens as proposed solution in financial data-to-text systems. An empirical study on the financial dataset validates our initial hypothesis that the character-based representation performs comparable results in content selection and diversity towards the generated text descriptions.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Natural Language Generation, Data-to-Text
Subjects: Q Science > QA Mathematics > QA71-90 Instruments and machines > QA75.5-76.95 Electronic computers. Computer science
Divisions: Faculty of Computing and Informatics (FCI)
Depositing User: Ms Rosnani Abd Wahab
Date Deposited: 15 Oct 2021 02:17
Last Modified: 15 Oct 2021 02:17


Downloads per month over past year

View ItemEdit (login required)